Lets get start with TensorFlow

December 26, 2016 0 Comments
installation , Junaid , setup , tensorflow

If you are interested in Machine or Deep Learning then you are at the right place. Today I am going to share the basic, setup, guide and use of TensorFlow with all respective references that you need to know.

Lets get start with TensorFlow

What is TensorFlow?
TensorFlow is an open source machine learning or deep learning library developed by Google team consisting of top researchers and engineers. The main purpose of this contribution is to help researchers, engineers and data scientists to carry out any ML task easily.

How to setup TensorFlow?
In order to setup TensorFlow environment on your computer, you have to install several things. TF is supported by all major operating systems windows, macOS and Linux, I will be discussing all the stuff using Linux, I will prefer you to install it on Linux for better use and understanding in the future.

TensorFlow setup steps for Linux:
- Download Ubuntu 16 (You may want to use it as a virtual machine on VMWare or VIPbox)
- Install python 3.5 (It might already be installed)
- Install pip3 (It’s a package to manage python software and libraries)
- Install tensorflow
I will prefer you to follow all the steps else you might face several issues, especially with the versions of Ubuntu and python.

Additional Installation
Jupyter (IPython Notebook), its interactive python shell, can also be run as a web on localhost.
A detailed step by step installation can be found at setup tensorflow.

How to use?
TF library is supported by several programming languages including Python, C++, I will use Python in the future articles.
The good thing about tensorflow is that it can also be used with Hadoop and GPU using Cuda Toolkit.
To get familiar with a tiny snippet of tensorflow in python, here is code.

import tensorflow as tf
x = tf.Variable([1.0, 2.0])
z = 2 * x
sess = tf.Session()
init = tf.global_variables_initializer()
print("x = ", sess.run(x))
print("z = 2 * x = ", sess.run(z))

You can run it once you are done with the setup. It is here just for testing purpose, I will explain this code in the next article in detail.

To get more help visit tensorflow.org or its official github repository.